Celastrol Alleviates Corneal Stromal Fibrosis by Inhibiting TGF-β1/Smad2/3-YAP/TAZ Signaling After Descemet Stripping Endothelial Keratoplasty

Abstract

Purpose: The purpose of this study was to investigate the effect of celastrol (CEL) on corneal stromal fibrosis after Descemet stripping endothelial keratoplasty (DSEK) and its associated mechanism.

Methods: Rabbit corneal fibroblasts (RCFs) were isolated, cultured, and identified. A CEL-loaded positive nanomedicine (CPNM) was developed to enhance corneal penetration. CCK-8 and scratch assays were performed to evaluate cytotoxicity and the effects of CEL on the migration of RCFs. The RCFs were activated by TGF-β1 with or without CEL treatment, and then the protein expression levels of TGFβRII, Smad2/3, YAP, TAZ, TEAD1, α-SMA, TGF-β1, FN, and COLI were assessed by immunofluorescence or Western blotting (WB). An in vivo DSEK model was established in New Zealand White rabbits. The corneas were stained using H&E, YAP, TAZ, TGF-β1, Smad2/3, TGFβRII, Masson, and COLI. H&E staining of the eyeball was performed to assess the tissue toxicity of CEL at 8 weeks after DSEK.

Results: In vitro CEL treatment inhibited the proliferation and migration of RCFs induced by TGF-β1. Immunofluorescence and WB showed that CEL significantly inhibited the protein expression of TGF-β1, Smad2/3, YAP, TAZ, TEAD1, α-SMA, TGF-βRII, FN, and COL1 induced by TGF-β1 in RCFs. In the rabbit DSEK model, CEL significantly reduced the levels of YAP, TAZ, TGF-β1, Smad2/3, TGFβRII, and collagen. No obvious tissue toxicity was observed in the CPNM group.

Conclusions: CEL effectively inhibited corneal stromal fibrosis after DSEK. The TGF-β1/Smad2/3-YAP/TAZ pathway may be involved in the mechanism by which CEL alleviates corneal fibrosis. The CPNM is a safe and effective treatment strategy for corneal stromal fibrosis after DSEK.

Corneal endothelial diseases, such as Fuchs' dystrophy, posterior polymorphous dystrophy, and corneal endothelial dysfunction caused by glaucoma, intraocular surgery, and inflammation, are among the main factors affecting visual acuity. Descemet stripping endothelial keratoplasty (DSEK) is an effective treatment for corneal endothelial diseases.¹ However, corneal stromal fibrosis impedes visual recovery after DSEK.² At present, medical management of corneal fibrosis relies on the application of topical corticosteroids and mitomycin C.³ However, these drugs often have limited efficacy and long-term complications, such as cataracts, glaucoma, and corneoscleral melting.⁴ Therefore, more effective approaches to prevent or alleviate corneal fibrosis are needed. 

Fibrosis, defined as the excessive accumulation of extracellular matrix, is a common pathological process associated with wound healing in many organs.⁵ The corneal wound healing response involves a cascade of events. Ultimately, excessive deposition of extracellular matrix (ECM) leads to corneal stromal fibrosis.^6,7 Transforming growth factor-β1 (TGF-β1), which induces keratocyte transformation to myofibroblasts, plays a vital role in corneal fibrosis. TGF-β1/Smad2/3 pathway activity causes myofibroblasts to produce large amounts of disordered ECM to generate corneal scarring.^6,8 Yes-associated protein (YAP) and transcriptional coactivator with PDZ binding motif (TAZ) proteins, which bind to TEAD transcription factors, are essential for profibrotic cues induced by TGF-β in the microenvironment.⁹ Elevated YAP/TAZ levels correlate with local tissue fibrosis. Pharmacological inhibition or knockdown of YAP/TAZ prevents TGF-β-induced renal fibrosis and delays skin wound healing.^10,11 Likewise, selective YAP/TAZ inhibition reverses pulmonary fibrosis in mice.¹² Collectively, these studies suggest that targeting YAP/TAZ activity may hold promise for the treatment of corneal stromal fibrosis. 

Celastrol (CEL) is a pentacyclic tri-terpenoid derived from Tripterygium wilfordii Hook F. The potential biological activities of CEL in autoimmune diseases, inflammatory diseases, cancer, and obesity associated diseases have been demonstrated in recent decades.^13–16 Recent studies have also shown that CEL has antifibrotic effects in the heart, liver, and kidneys.^17–19 Divya et al. reported that CEL can reverse bleomycin-induced pulmonary fibrosis through the TGF-β1/Smad2/3 pathway in rats.²⁰ CEL, a new inhibitor of the YAP/TAZ interaction, can inhibit cancer cell proliferation, transformation, and cell migration.²¹ Whether CEL inhibits corneal stromal fibrosis through TGF-β1/Smad2/3-YAP/TAZ signaling has not been reported. In this study, we tried to elucidate the effect of CEL on corneal stromal fibrosis after DSEK. 

Although CEL has potential as an antifibrosis drug, its low solubility and poor bioavailability limit its clinical application.^22,23 To improve the solubility of CEL, we have developed new drug delivery systems for ophthalmic administration.^24,25 Nevertheless, the ocular bioavailability of the instilled drugs is very low (less than 5%) because of the rapid renewal rate of the lymphatic fluid together with the blinking reflex.²⁶ In addition, the highly organized multilayered and lipophilic corneal epithelium and hydrophilic stroma make drug transport very difficult. To overcome the issue of low bioavailability and enhance corneal penetration, we developed a CEL-loaded positive nanomedicine (CPNM) for possible application in the cornea.²⁷ In the present study, we observed fibrosis-related protein generation after TGF-β1 stimulation in rabbit corneal fibroblasts (RCFs). Then, we evaluated the effect of the CPNM on corneal stromal fibrosis after DSEK in rabbits. 

Methods

Antibodies

Antibodies against Vimentin (ab45939), Lumican (bs-5890R), Keratocan (bs-11054R-PE), α-SMA (ab7817), YAP (bs-3605R), TAZ (ab84927), TGFβRII (ab186838), TGF-β1 (AF1027), TEAD1 (12292S), P-Smad2/3 (AF3367), FITC-labeled secondary antibodies (ab6717), a fluorescent secondary antibody (Cy3) (bs-0295P-PE-Cy3), β-actin (20536-1-AP-1), β-tubulin (10094-1-AP) and GAPDH (5174S); and FITC-labeled secondary antibodies (ab6717), a fluorescent secondary antibody (Cy3) (bs-0295P-PE-Cy3). 

Preparation and Characterization of the CPNM

The copolymer poly(ethylene glycol)-poly(ε-caprolactone)-g-polyethyleneimine (PCI) was synthesized as previously described.²⁵ The methods used to prepare and characterize CPNM were described in our previously published article.²⁷ 

Extraction and Culture of Primary Rabbit Corneal Fibroblasts

The RCFs were isolated from corneas. In detail, freshly isolated corneas were immediately soaked in 0.5% gentamicin for 2 hours. The stromal tissue was divided into small pieces and inoculated into culture bottles in DMEM/F12 medium with 10% fetal bovine serum (FBS). When the cells reached approximately 60% confluence, they were subpassaged. Cells not exceeding passage five were used in the experiments. 

RCF Identification

The F1 RCFs were incubated with primary antibodies against Vimentin (1:200, ab45939), Lumican (1:100, bs-5890R), and Keratocan (1:200, bs-11054R-PE) at 4°C overnight. Immunoreactivity was evaluated using two fluorescein isothiocyanate (FITC)-labeled secondary antibodies (Abcam Inc., Cambridge, MA, USA) and a fluorescent secondary antibody (Cy3; Bioss, Beijing, China), and cell nuclei were counterstained with Fluoroshield containing 4′-6-diamidino-2-phenylindole (DAPI; Sigma F6057-20ML). Preparations were evaluated using fluorescence microscopy (Nikon 80i). RCFs were identified by flow cytometry to detect Vimentin according to the manufacturer's protocol. 

CPNM Cellular Uptake Test

The cellular uptake test was performed at 1 hour, 2 hours, and 4 hours after the F2 RCFs were cultured with CPNM (CEL, 4 µg/mL). Based on the spontaneous fluorescence of CEL micelles, fluorescence microscopy images were semiquantitatively analyzed with GraphPad software to evaluate the uptake of CEL micelles by RCFs. 

Cell Cytotoxicity and Scratch Assays

To investigate the cytocompatibility of CEL in the positive nanomedicine (PNM), the Cell Counting Kit-8 assay (CCK-8; Dojindo) was utilized. Briefly, after treatment with different concentrations of the PNM (0.75, 1.5, 3, 6, 12, and 24 µg/mL) for 24 hours or the PNM at 24 µg/mL for 48 hours, 72 hours, and 96 hours, the RCFs were cultured in CCK-8 solution for 4 hours at 37°C in 96-well plates, and a microplate reader was used to test the absorbance of the supernatant 450 nm. 

To evaluate the effect of CEL on the proliferation of RCFs induced by TGF-β1, RCFs were cultured in medium containing 10 ng/mL TGF-β1 for 24 hours and then treated with different concentrations of the CPNM (0.05, 0.1, 0.2, 0.4, 0.8, and 1.6 µg/mL) for 24 hours. Then, cell survival was quantitatively evaluated using a CCK-8 kit. 

Scratch assays were performed to detect RCF migration after TGF-β1 and CPNM treatment. Briefly, when cells grown in 6-well plates reached 80% confluence, the RCFs were treated with TGF-β1 (10 ng/mL) for 24 hours and starved for 8 hours in DMEM/F12 medium containing 0.5% FBS. The RCFs were scratched with a pipette tip and rinsed two to three times. After treatment with the CPNM (0.05, 0.1 µg/mL) for 12 hours, the migrated cells were counted. The percentage migration was quantified with the percentage migration in untreated wells set to 100%. 

Immunofluorescence Assay

The RCFs were pretreated with CEL (0.1 µg/mL) for 4 hours, and the medium containing CEL was discarded. Then, the RCFs were induced by TGF-β1 (10 ng/mL) treatment for another 24 hours. The RCFs were incubated with primary antibodies against: α-SMA (1:200, ab7817), YAP (1:200, bs-3605R), TAZ (1:200, ab84927), TGFβRII (1:200, ab186838), TGF-β1 (1:200, AF1027), TEAD1 (1:200, 12292S), FN (1:400, NBP1-51723), COLI (1:400, NB600-408), and P-Smad2/3 (1:200, 12292S) at 4°C overnight. Immunoreactivity was evaluated using two FITC-labeled secondary antibodies, and cell nuclei were counterstained with DAPI. The preparations were evaluated using fluorescence microscopy (Nikon 80i). The fluorescence microscope images were analyzed semiquantitatively by ImageJ and GraphPad software. 

Western Blot Analysis

RCFs were seeded on sterile cover slides and cultured for 24 hours in 6-well plates. The RCFs were pretreated with CEL (0.1 µg/mL) for 4 hours. The CEL-containing medium was discarded. The RCFs were induced by TGF-β1 (10 ng/mL) treatment for another 24 hours. Total protein was extracted from the cells and sonicated in RIPA buffer (C1053; Applygen, Beijing, China). A BCA assay reagent kit (P1511; Applygen) was used to measure the protein concentration. The proteins were transferred to an Immun-Blot polyvinylidene difluoride (PVDF) membrane that was subsequently blocked and probed with primary antibodies against one of the following: α-SMA (1:1000, ab7817), YAP (1:500, bs-3605R), TAZ (1:1000, ab84927), TGF-βRII (1:1000, ab186838), TGF-β1 (1:1000, AF1027), TEAD1 (1:1000, 12292S), P-Smad2/3 (1:1000, AF3367), β-actin (1:500, 20536-1-AP-1), β-tubulin (1:500, 10094-1-AP), and GAPDH (1:500, 5174S). The membranes were then incubated with HRP-conjugated secondary antibodies. Protein bands were visualized using Amersham Biosciences ECL Western blot detection reagent. The protein bands were analyzed semiquantitatively by ImageJ and GraphPad software. 

Animals

Adult New Zealand white rabbits (12 weeks, 2.5–3.5 kg) were used as the animal DSEK model. The experimental protocol was approved by the Committee on the Ethics of Animal Experiments of Henan Eye Hospital (Permit number: HNEECA-2022-19). Slit lamp examinations showed that no eyes had any corneal opacity or imperfections. All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

DSEK Model

DSEK surgery was performed according to the standard DSEK procedure applied in clinical settings.²⁸ Briefly, for the DSEK graft, 300 µm lamellar keratectomy was performed on donor rabbit corneas using a depth knife (MDP 30; Mani Inc., Japan) and a straight 45 degree knife (MST 45; Mani Inc., Japan), and the residual corneal bed was marked with F and trephined with a 6-mm-diameter trephine. In the recipient rabbit, which was under general anesthesia, the Descemet membrane was stripped from recipient corneas. DSEK grafts were transplanted using the standard DSEK procedure. The operative incision was closed with interrupted 10.0 nylon sutures (Alcon, Fort Worth, TX, USA). All eyes were administered Tobradex eye drops (tobramycin 0.3% and dexamethasone 0.1%, 5 mL; Alcon, USA) four times a day and TobraDex eye ointment (tobramycin 0.3% and dexamethasone 0.1%, 3.5 g; Alcon, USA) once a night for 2 weeks as prophylaxis. Dil labeling was used to trace the implanted DSEK graft. The DSEK graft was incubated in Dil (12.5 mg/mL) for 10 minutes and then was grasped and transplanted according to the DSEK procedure. 

Dil Staining

Dil (1,1′-dioctadecyl-3-3-3′, 3′-tetramethylindocarbocyanine; Molecular Probes; Sigma-Aldrich) labeling was used to trace the implanted DSEK graft. The DSEK graft was incubated in Dil (12.5 mg/mL) for 10 minutes and then was grasped and transplanted according to the DSEK procedure. Following removal of the eyeball at 2 weeks, the fresh cornea was immediately flash frozen in liquid nitrogen. The frozen tissue sections were fixed in 4% paraformaldehyde and stained with DAPI. Fluorescence images were recorded using fluorescence microscopy (Nikon 80i). 

Experimental Design

A total of 20 rabbits were divided randomly into two groups (n = 10 per group): the control group (PNM group) and the CPNM group. The PNM or CPNM (10 µL at a time) was administered via eye drops 2 times a day for a total of 8 weeks until the animals were euthanized. All rabbits were subjected to routine postoperative slit lamp examination daily for 2 weeks and then weekly for the next 6 weeks. The rabbits were euthanized by administration of an overdose of pentobarbital sodium 2 weeks and 8 weeks after DSEK. 

Histology

Corneal tissues were collected and embedded in paraffin, followed by slicing into sections at 2 and 8 weeks. H&E and Masson staining was performed to detect changes in the cornea in the two groups. Immunohistochemical staining, including staining of YAP, TAZ, and COLI, was conducted according to the manufacturer's protocol. Immunofluorescence staining, including P-Smad2/3, TGF-β1, and TGFβRII, was conducted according to the manufacturer's protocol. H&E staining of the eye tissue structure was conducted to assess the tissue toxicity of the CPNM at 8 weeks after DSEK. The stained sections were observed using an optical microscope (XSP-C204, CIC), and images were captured with a 3D digital slice scanner (3DHISTECH P250). 

Statistical Analysis

All experiments were repeated three times. In addition, the results are expressed as the mean ± SD. Differences between data sets were evaluated for statistical significance using Student's t-test or one-way analysis of variance (SPSS 25.0, Inc., Armonk, NY, USA). P < 0.05 was considered significant. 

Results

Biocompatibility of the PNM and Cellular Uptake of the CPNM

To identify RCFs, we applied Vimentin, Lumican, and Keratocan immunofluorescence staining. The cells had a star-like or spindle-shape with appearance of an oval-shaped nucleus in the center of each cell (Fig. 1A). Vimentin⁺, Lumican⁺, and Keratocan⁺ staining confirmed that the cells cultured in vitro were RCFs. Further identification of RCFs by flow cytometry to detect Vimentin was performed, and the percentage of Vimentin-labeled RCFs was 99% (Fig. 1B). We used RCFs to examine the cytocompatibility of the PNM by CCK-8 assay. The proliferation of RCFs in the PNM group was similar to that in the control group (Fig. 1C, cell activity >90%), which indicated that there was no difference in cell activity among cells treated with different concentrations of the PNM or for different durations. 

Figure 1.

 

Identification of RCFs, biocompatibility of the PNM, and cellular uptake assay of the CPNM. (A) Immunofluorescence of vimentin (VIM, green), lumican (green), and keratocan (red) (400 ×). Bar = 20 µm. DAPI, (4′-6-diamidino-2-phenylindole, blue). (B) The rate of VIM-labeling was 99% by flow cytometry in RCFs. (C) The cytocompatibility of the PNM by CCK-8 assay at different concentrations for 24 hours or the PNM at 24 µg/mL for different durations (n = 3). (D) Cellular uptake and intracellular distribution of the CPNM (green, 1000 ×) at different time points. Bar = 20 µm. (E) Mean fluorescence intensity of the CPNM at 1 hour, 2 hours, and 4 hours after incubation with the CPNM (n = 3, CEL concentration = 4 µg/mL). ***P < 0.001.

Due to the spontaneous fluorescence of CEL, cell uptake behavior could be directly observed under a fluorescence microscope. As shown in Figures 1D and 1E, the cellular uptake of the CPNM by the RCFs gradually increased over 4 hours. 

Effects of the CPNM on Cell Proliferation and Migration Induced by TGF-β1

We evaluated the effect of CEL on the proliferation of RCFs induced by TGF-β1 with CCK-8. Our results demonstrated that the cell viability was 79.71% when the CPNM concentration was 0.1 µg/mL, 75.21% when it was 0.2 µg/mL, 64.33% when it was 0.4 µg/mL, 52.97% when it was 0.8 µg/mL, and 28.93% when it was 1.6 µg/mL, as shown in Figure 2A. This suggested that CEL inhibited RCF proliferation in a concentration-dependent manner. However, a high concentration of CEL (1.6 µg/mL) led to a decrease in adhesion ability and even death of RCFs in subsequent culture. 

Figure 2.

 

CEL inhibited the proliferation and migration of RCFs induced by TGF-β1 in vitro. (A) Cell proliferation was determined by CCK-8 assay with the CPNM at different concentrations. (B) Representative micrographs of RCF migration induced by TGF-β1 for 12 hours in the scratch assay (1000 ×). Scale bar = 100 µm. (C) Quantitative analysis of migration; n = 3, *P < 0.05 **P < 0.01 ***P < 0.001.

In the present study, we observed the effect of the CPNM on the number of TGF-β1-induced RCFs that migrated. As shown in Figures 2B and 2C, the scratch assay results showed that the number of migrating cells was higher in the control group than in the CPNM group (0.05 and 0.1 µg/mL). This result indicated that the migration capacities of the cells after CEL treatment were significantly decreased. 

CEL Reduced Fibrosis-Related Protein Expression in RCFs

To assess the effect of CEL on RCFs induced by TGF-β1, fibrosis-related proteins were assessed by immunofluorescence staining or Western blotting (WB). Our results showed that the fluorescence levels after staining of the TGF-βRII, YAP, TAZ, TEAD1, and P-Smad2/3 proteins in the TGF-β1+CEL group were significantly lower than those in the TGF-β1 group (P < 0.001; Fig. 3). In addition, there was nearly no difference in average fluorescence intensity except for TEAD1 and P-Smad2/3 between the control group and the TGF-β1+CEL group (see Fig. 3). The fluorescence levels after staining of the TGF-β1, α-SMA, FN, and COLI proteins in the TGF-β1+CEL group were significantly reduced compared with those in the TGF-β1 group (P < 0.001; Fig. 4). In addition, there was no significant difference in the mean fluorescence intensity (MFI), including for TGF-β1 and α-SMA, between the control group and the TGF-β1+CEL group (see Fig. 4). The levels of the TGF-βRII, TGF-β1, YAP, TAZ, TEAD1, P-Smad2/3, and α-SMA proteins significantly increased in the TGF-β1 group compared with the TGF-β1+CEL group (P < 0.001; Fig. 5). The levels of the TGF-β1, YAP, TEAD1, and α-SMA proteins decreased significantly in the TGF-β1+CEL group compared with the control group (P < 0.01 and P < 0.001; see Fig. 5). 

Figure 3.

 

Expression of fibrosis-related proteins in RCFs induced by TGF-β1 as shown by immunofluorescence. (A) Representative micrographs of TGF-βRII, YAP, TAZ, TEAD1, and P-Smad2/3 (green, 400 ×) in RCFs. DAPI staining is blue in all panels. Scale bar = 20 µm. (B) The mean fluorescence intensity of fibrosis-related proteins. Data are presented as the mean ± SD; n = 3 per group, *P < 0.05, ***P < 0.001.

Figure 4.

 

Immunofluorescence for TGF-β1, α-SMA, FN, and COLI in RCFs activated by TGF-β1. (A) Representative immunofluorescence micrographs of fibrosis-related proteins. Green indicates TGF-β1 (400 ×), α-SMA (400 ×), FN (1000 ×), and COLI (1000 ×) staining. Blue is DAPI staining in all panels. Scale bar = 20 µm. (B) The mean fluorescence intensity (MFI) of fibrosis-related proteins. Data are presented as the mean ± SD; n = 3 per group, *P < 0.05, ***P < 0.001.

Figure 5.

 

Confirmation by WB of the inhibitory effect of CEL on fibrosis-related proteins in RCFs activated with TGF-β1. (A) Representative immunoblotting images are shown. (B) Analysis of relative protein expression. The expression of fibrosis-related proteins decreased significantly with CPNM treatment (**P < 0.01). Data are presented as the mean ± SD; n = 3 per group, **P < 0.01, ***P < 0.001.

Evaluation of the Antifibrotic Effects of the CPNM In Vivo

We next determined the effect of the CPNM on corneal stromal fibrosis in vivo. Follow-up examination of rabbit eyes was carried out at 1 day, 1 week, 2 weeks, 4 weeks, 6 weeks, and 8 weeks after DSEK. On the first day after DSEK, the corneal graft bed and DSEK graft both showed mild edema. Figure 6A shows attachment of the graft to the back side of the cornea after DSEK. The cornea appeared transparent under the diffuse light of the slit lamp. There were no serious complications, such as cataracts, graft rejection, graft failure, and graft dislocation. However, corneal intrastromal opacity in the control groups was more obvious than that in the CPNM groups under the light of the slit lamp (see Fig. 6A). These results suggested that CEL can reduce the formation of interlamellar opacities during stromal wound healing and may improve postoperative visual acuity after DSEK. The endothelial cells of corneal grafts were labeled with Dil, and frozen sections were generated 2 weeks after DSEK (Fig. 6B). The results showed that the grafts survived after DSEK. H&E staining demonstrated that there was no significant infiltration of immune cells into the corneal endothelium of a DSEK graft in the 2 groups at 2 weeks and 8 weeks (Fig. 6C). 

Figure 6.

 

Slit-lamp photographs and representative H&E staining images. (A) Slit-lamp photographs of the cornea 1 day, 1 week, 2 weeks, 4 weeks, and 8 weeks after DSEK. (B) Dil-labeled endothelial cells (red) were observed in DSEK grafts 2 weeks after DSEK. Scale bar = 20 µm. (C) H&E staining images from the control and CPNM groups at 2 weeks and 8 weeks after DSEK; n = 3 per group, Scale bar = 500 µm for 2 ×, and 20 µm for 40 ×.

Next, the immunohistochemical staining of YAP/TAZ was evaluated at 2 weeks after DSEK. In addition, the immunofluorescence staining of TGF-β1/TGF-βRII/Smad2/3 was evaluated at 2 weeks after DSEK. Figures 7A and 7B show lower expression of YAP and TAZ in the corneal stroma in the CPNM group. Figures 7C and 7D show lower expression of TGF-β1, TGF-βRII, and P-Smad2/3 in the corneal stroma in the CPNM group. In addition, COLI staining and Masson staining were used to evaluate the synthesis of collagen fibers at 8 weeks after DSEK. The results showed that collagen formation in the stromal layer of the cornea in the CPNM group was hardly observed compared with that in the control group (see Figs. 7A, 7B, P < 0.001). 

Figure 7.

 

Evaluation of the antifibrotic effects of the CPNM in vivo. (A) Immunohistochemistry for YAP, TAZ, and COLI and Masson staining images at 2 weeks and 8 weeks after DSEK. Scale bar = 20 µm. (B) Semiquantitative analysis. Data are presented as the mean ± SD; n = 3 per group, ***P < 0.001. (C) Immunofluorescence for TGF-β1, TGF-βRII, and P-Smad2/3 staining images at 2 weeks after DSEK. Scale bar = 100 µm. (D) The mean fluorescence intensity. Data are presented as the mean ± SD; n = 3 per group, ***P < 0.001.

Biological Safety of CPNM

H&E staining was performed to assess the safety of the CPNM at 8 weeks after DSEK. No significant differences were found in eye tissues, including the cornea, conjunctiva, iris, ciliary body, and retina, among the normal, control, and CPNM groups (Fig. 8). 

Figure 8.

 

The tissue safety assessment of the CPNM at 8 weeks after DSEK. H&E staining, including cornea, conjunctiva, iris, ciliary body, and retina in the normal, control, and CPNM groups; n = 3 per group. Scale bar = 20 µm.

Discussion

After surgery, traumatic injury, or infection-related injury, corneal wound healing can cause corneal stromal fibrosis and vision loss.²⁹ Fibrosis between the corneal stromal layers after DSEK is a key factor affecting the recovery of visual acuity. Based on our results, CEL inhibited myofibroblast differentiation and reduced corneal fibrosis after DSEK. Thus, CEL is a potential therapeutic strategy and may provide new ideas or drugs for corneal stromal fibrosis after DSEK. 

In the present research, the CPNM had good cytocompatibility, which was consistent with our previous research results.²⁷ Moreover, we observed no obvious abnormal ocular inflammatory or toxic reactions 2 months after DSEK. In our previous study, we performed a local eye irritation test. We found that the CPNM was not an irritant and could be tolerated by the rabbit eye.²⁷ Taken together, these data reveal that the PNM has good biocompatibility, which suggests that the CPNM has potential for future clinical application. 

Several signaling pathways are involved in fibrosis, including TGF-β, Wnt, and Rho, which interact with YAP/TAZ in pathological fibrosis.^30,31 TGF-β 1/Smad2/3 is a classical signal pathway in the process of tissue fibrosis. In addition, TGF-β1 is also the most important fibrogenic and growth-regulating cytokine in corneal wound healing.³ TGF-β1 exerts its effect via the heterodimerization of TGF-β receptor type I (TGF-βRI) and TGFβRII.³² YAP/TAZ-TEAD are transcription cofactors that activate TGF-β1/Smad2/3 in the nucleus.¹⁰ YAP/TAZ binding to the TEAD transcription factors has been linked to most pathophysiological process, such as wound healing, organ development, epithelial homeostasis, tissue regeneration, and immune modulation.⁹ YAP/TAZ promoted myofibroblast differentiation and increased matrix remodeling potential in fibroblasts of the liver,³¹ kidneys,³³ lungs,³⁴ and skin.³⁵ YAP/TAZ converge with pro-fibrotic TGF-β1 signaling pathways. YAP/TAZ are indeed essential for pro-fibrotic process induced by TGF-β1.³⁶ Muppala et al. also reported that YAP/TAZ played critical roles in TGF-β1 induced myofibroblast transformation in primary human corneal fibroblasts.³⁷ In the present study, our results demonstrated an increase in YAP, TAZ, TEAD1, α-SMA, FN, and COL1 expression after TGF-β1 stimulation in vitro. These data suggested that YAP/TAZ-TEAD may interact with TGF-β1/Smad2/3 signaling and mediate nuclear translocation during corneal fibroblast activation. In vivo, the expression of TGF-β1, P-Smad2/3, YAP, and TAZ increased in the cornea in the early stage after DSEK in the control group. These results suggested that the TGF-β1/Smad2/3-YAP/TAZ signaling pathway is involved in the process of myofibroblast activation and corneal fibrosis after DSEK. 

Activation of YAP/TAZ promoted fibrosis in multiple organs, including the skin, liver, kidneys, and lungs.³⁸ Selective inhibition of YAP/TAZ has been shown to attenuate kidney and lung fibrosis.^12,39 This finding suggests the relevance of YAP/TAZ as a core pathway in fibrosis. In the present study, we found that corneal fibroblasts were unresponsive to TGF-β1 stimulation after CEL treatment. CEL induced a dramatic reduction in YAP/TAZ-TEAD1 and TGF-β1/Smad2/3 expression in vitro. Our data also showed decreased expression of α-SMA and ECM proteins, such as FN and COL1, in the CEL group. In addition, CEL inhibited proliferation and migration induced by TGF-β1 in RCFs. These data demonstrated that CEL may suppress corneal fibroblast activation via TGF-β1/Smad2/3-YAP/TAZ signaling. In vivo, we also found that CEL, a YAP/TAZ inhibitor, decreased the expression of TGF-β1, P-Smad2/3, and YAP/TAZ and reduced collagen deposition in the cornea after DSEK. This suggests that CEL may alleviate corneal stromal fibrosis by inhibiting TGF-β1/Smad2/3-YAP/TAZ signal. 

To the best of our knowledge, this is the first study to demonstrate the role of CEL in corneal stromal fibrosis after DSEK. We infer from these in vitro and in vivo findings that CEL may reduce corneal fibrosis by inhibiting the TGF-β1/Smad2/3-YAP/TAZ pathway. However, there are still some limitations of this study. Because CEL has a potent biological effect, there may be other mechanisms affecting the biological function of corneal fibroblasts. Further work is needed to elucidate the associations between other molecular mechanisms and the YAP/TAZ pathway. Further research should be done, including the expressions of protein and RNA in corneal tissue, to demonstrate that CEL can alleviate corneal stromal fibrosis through TGF-β1/Smad2/3-YAP/TAZ signal at multiple levels. In addition, TGF-β1, Smad2/3, and YAP/TAZ inhibitors should be added to further clarify the role of TGF-β1/Smad2/3-YAP/TAZ in corneal stromal fibrosis. 

Conclusion

Our findings suggest that CEL can inhibit myofibroblast differentiation, impede corneal fibroblast proliferation, and alleviate corneal stromal fibrosis after DSEK. Our study revealed that CEL may exert its antifibrotic effects in the cornea via TGF-β1/Smad2/3-YAP/TAZ signaling. CEL may be used as an antifibrotic agent for corneal fibrosis treatment. These results reveal a safe and effective strategy for anti-scarring treatment after DSEK. 

Acknowledgments

Supported by the National Natural Science Foundation of China (82101090 and 52173143), Project of Henan Provincial Health Department (200803096); Medical Science and Technology Project of Henan Province (SBGJ202103012), and Basic Science Key Project of Henan Eye Institute/Henan Eye Hospital (20JCZD002, 20JCQN001, and 22JCQN001). The authors are alone responsible for the content and writing of the paper. 

Disclosure: R. Liu, None; J. Li, None; Z. Guo, None; D. Chu, None; C. Li, None; L. Shi, None; J. Zhang, None; L. Zhu, None; Z. Li, None